Ceph

01/03/2023

08:01 AM bluestore Bug #58274: BlueStore::collection_list becomes extremely slow due to unbounded rocksdb iteration

yixing hao wrote:
> Also observed from our HDD bluestore cluster with tens of billions of objects, the stack is like... yixing hao

07:51 AM bluestore Bug #58274: BlueStore::collection_list becomes extremely slow due to unbounded rocksdb iteration

Also observed from our HDD bluestore cluster with tens of billions of objects, the stack is like the above.
7ffad9... yixing hao

11/11/2022

09:14 AM CephFS Bug #58008: mds/PurgeQueue: don't consider filer_max_purge_ops when _calculate_ops

When increasing filer_max_purge_ops on a pacific version mds, pq_executing_ops/pq_executing_ops_high_water of purge_q... yixing hao

09:13 AM CephFS Bug #58008 (Resolved): mds/PurgeQueue: don't consider filer_max_purge_ops when _calculate_ops

_calculate_ops relying on a config which can be modified on the fly will cause a bug. e.g.
# A file has 20 objects... yixing hao

06/15/2022

08:55 AM CephFS Feature #56058: mds/MDBalancer: add an arg to limit depth when dump loads for dirfrags

https://github.com/ceph/ceph/pull/46685 yixing hao

08:54 AM CephFS Feature #56058 (Pending Backport): mds/MDBalancer: add an arg to limit depth when dump loads for dirfrags

Directory hierarchy may be deep for a large filesystem, cmd dump loads would output
a lot and take a long time. So a... yixing hao

03/30/2022

11:43 AM CephFS Feature #55126: mds: add perf counter to record slow replies

Reviewer suggested that the pr could be backported, so create the tracker here. yixing hao

11:42 AM CephFS Feature #55126: mds: add perf counter to record slow replies

Though we have MDS_HEALTH_SLOW_METADATA_IO and MDS_HEALTH_SLOW_REQUEST health alert, but those are not
precise nor a... yixing hao

11:42 AM CephFS Feature #55126 (Pending Backport): mds: add perf counter to record slow replies

yixing hao